Method and system for realizing human motion state recognition through radar

Abstract

The invention relates to the technical field of radar motion recognition, in particular to a method and system for recognizing a human motion state through radar, and the method comprises the following steps: receiving echo signals of millimeter wave radar through a sensor, analyzing the amplitude and phase change of the signals, calculating the average value and standard deviation of the amplitude, and comparing the received signals. And recording an amplitude extreme point and a phase abrupt change point to obtain a radar echo signal feature. According to the method, the amplitude and phase change of the radar echo signals are meticulously analyzed, the execution process optimizes the monitoring of the motion state of the human body, while key signal features are extracted, the amplitude average value and the standard deviation of the signals are calculated in a continuous time period, and the sensitivity and the accuracy of signal processing are further improved; by analyzing the gradient value of the phase change and comparing the change of the adjacent time points, the method effectively recognizes the fine change of the human body movement speed, and greatly improves the capturing capability of the dynamic movement track.

Description

Technical Field

[0001] The present invention relates to the technical field of radar motion recognition, and particularly to a method and system for realizing human motion state recognition through radar. Background Art

[0002] The technical field of radar motion recognition includes using a radar system to detect and recognize the motion state of an object, especially a human body, in space. This technical field is mainly based on the transmission and reception of radar waves, and by analyzing the changes in the reflected waves, dynamic information such as the position, speed, and their changes of the object can be obtained. Radar motion recognition technology is widely used in many fields such as security monitoring, health monitoring, and interactive entertainment. The core contents include the transmission modulation of radar waves, the capture of echo signals, signal processing technology, and the parsing algorithm of the motion state.

[0003] Among them, the method for realizing human motion state recognition through radar refers to using a specific type of radar device to emit electromagnetic waves and capturing the changes in the electromagnetic waves caused by human motion. The technical matters covered include the selection of radar waves, the design of transmitters and receivers, and the signal processing process for parsing human motion characteristics. The specific method is to recognize specific actions or the overall motion state of the human body by analyzing the radar echoes, involving time-frequency analysis of radar signals to extract motion information.

[0004] The existing technology has insufficient capture of rapid or minute action changes, especially in action recognition in a multi-person environment and precise measurement of changes in motion speed. There are obvious deficiencies in fields such as security monitoring and health monitoring that require high-precision monitoring, resulting in the inability to meet actual application requirements. For example, quickly recognizing specific actions in emergency situations or maintaining high-accuracy motion tracking in complex backgrounds becomes a challenge, leading to misjudgment or missed detection in actual operation, affecting the overall performance and reliability of the system. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, such as insufficient capture of rapid or minute action changes, especially in action recognition in a multi-person environment and precise measurement of changes in motion speed, resulting in the inability to meet actual application requirements in fields such as security monitoring and health monitoring that require high-precision monitoring. For example, quickly recognizing specific actions in emergency situations or maintaining high-accuracy motion tracking in complex backgrounds becomes a challenge, leading to misjudgment or missed detection in actual operation, affecting the overall performance and reliability of the system, the embodiments of the present invention provide a method and system for realizing human motion state recognition through radar. The technical solutions are as follows:

[0006] On the one hand, a method for realizing human motion state recognition through radar is provided, including the following steps:

[0007] S1: Use a sensor to receive the echo signal of the millimeter-wave radar, analyze the amplitude and phase changes of the signal, calculate the average value and standard deviation of the amplitude, compare the received signal, record the amplitude extreme points and phase mutation points, and obtain the radar echo signal characteristics;

[0008] S2: Based on the radar echo signal characteristics, segment the phase information of the signal, calculate the phase change gradient value, analyze the gradient stable interval, identify the human body movement trajectory, judge the change of the movement speed, and obtain the movement mode data;

[0009] S3: Based on the movement mode data, extract the time points when the signal phase and amplitude are abnormal, identify the signal frequency components, compare the frequency and energy characteristics of the signal, judge the frequency mode of the movement state, identify the key frequency characteristics, and obtain the key frequency identification result;

[0010] S4: Based on the key frequency identification result, analyze the spectrum morphology difference, match the known frequency mode, identify the spectrum characteristics during movement, classify the movement mode, and obtain the movement state classification result;

[0011] S5: Based on the movement state classification result, calculate the movement change rate, analyze the violently fluctuating movement mode, identify the change of the echo signal intensity, judge the abnormal situation of the human body movement, and obtain the abnormal movement state index.

[0012] On the other hand, the radar echo signal characteristics include amplitude extreme points, phase mutation points and signal change amplitude, the movement mode data includes human body movement trajectory, movement speed change rate and phase stable interval, the key frequency identification result includes dominant frequency change, energy distribution characteristics and frequency component difference, the movement state classification result includes movement mode matching degree, spectrum morphology difference and movement state category, and the abnormal movement state index includes static or accelerating abnormal situation, multi-channel signal intensity change, movement state change rate.

[0013] On the other hand, the steps for obtaining the radar echo signal characteristics are specifically as follows:

[0014] S101: Use a sensor to receive the echo signal of the millimeter-wave radar, analyze the signal amplitude and phase information, calculate the signal amplitude mean value and standard deviation within a continuous time period, compare the received real-time signal, identify the signal amplitude mutation interval, calculate the signal energy change rate, and obtain the signal amplitude change characteristics;

[0015] S102: Based on the signal amplitude change characteristics, analyze the phase change of the signal, compare the phase offset amount of adjacent time points, analyze the continuity of the phase change, identify the cumulative offset trend of the phase change, judge the echo signal state, and obtain the phase offset characteristics;

[0016] S103: Invoke the phase offset characteristic to synchronize the signal segment with prominent signal amplitude and phase changes, calculate the change rate of signal energy before and after the mutation point, determine the key change positions in the echo signal, analyze the extreme points of the echo signal amplitude and the phase turning points, and obtain the radar echo signal characteristics.

[0017] On the other hand, the steps for obtaining the motion mode data are specifically as follows:

[0018] S201: Based on the radar echo signal characteristics, segment the phase information in the continuous echo signal, calculate the phase change amount, screen the signal intervals with abnormal phase changes, calculate the average phase increment in the interval, and compare the phase change rates of each signal segment to obtain the phase change rate index;

[0019] S202: Based on the phase change rate index, calculate the signal phase amplitude range in the continuous time period, compare the phase change directions of adjacent time windows, analyze the continuous change characteristics of the signal over time, identify the stable motion path of the human body, and obtain the stable motion trajectory;

[0020] S203: Invoke the stable motion trajectory, analyze the direction of human motion, screen the time nodes with changes in the motion trajectory direction, judge the amplitude and duration of the trajectory direction adjustment, compare the motion rates of adjacent trajectory segments, identify the change in the motion rhythm of the human body, record the change intervals of the motion direction and speed, and obtain the motion mode data.

[0021] On the other hand, the average phase increment in the calculation interval adopts the formula:

[0022]

[0023] And compare the phase change rates of each signal segment to obtain the phase change rate index;

[0024] Among them, represents the average phase increment in the signal interval, N represents the total number of sampling points in the selected signal interval, represents the phase increment of the i-th sampling point, represents the phase standard deviation of the i-th sampling point, represents the phase mean of the i-th sampling point.

[0025] On the other hand, the steps for obtaining the key frequency recognition result are specifically as follows:

[0026] S301: Based on the motion mode data, detect the phase and amplitude fluctuations of the echo signal, screen the time points with prominent change rates, calculate the amplitude difference of the continuous signals before and after the time point, analyze the change law of the phase mutation, measure the duration of the signal fluctuation, screen the signal feature points with amplitude and phase mutations, and obtain the signal fluctuation feature set;

[0027] S302: Call the signal fluctuation feature set, extract the echo signals at corresponding time points, analyze the amplitude changes of the signals in different frequency bands, identify the frequency composition components in the signals, calculate the rate of change of frequency over time, judge the frequency offset trend between time points, analyze the frequency distribution within adjacent time windows, and obtain a frequency offset map;

[0028] S303: Based on the frequency offset map, analyze the energy density of each frequency interval, compare the energy proportion of each frequency band, filter out the stable frequency regions, identify the frequency characteristics in the normal motion state, judge the matching relationship between the motion states of key frequency patterns, and obtain the key frequency identification result.

[0029] On the other hand, the steps for obtaining the motion state classification result are specifically as follows:

[0030] S401: Based on the key frequency identification data, analyze the spectral shape of the signals during human motion, extract the signal characteristics corresponding to each frequency pattern, compare the morphological differences of the signal waveforms, identify the spectral changes of the signals in each motion state, and obtain a spectral difference identification result;

[0031] S402: Call the spectral difference identification result, compare with the known frequency patterns, filter out the spectral shapes that meet the pattern characteristics, calculate the energy distribution ratio of each spectrum in the signal, analyze the matching degree between each pattern, determine the corresponding relationship between each spectral shape and the human motion state, and obtain motion spectrum matching data;

[0032] S403: Call the motion spectrum matching data, analyze the proportion of the motion mode signals, filter out the motion modes with significant proportions, analyze the time persistence of the corresponding modes, compare the signal frequency changes between different motion states, judge the motion category to which the echo signal belongs, determine the motion mode classification interval, and obtain the motion state classification result.

[0033] On the other hand, when calculating the energy distribution ratio of each spectrum in the signal, the formula is used:

[0034]

[0035] Analyze the matching degree between each pattern, determine the corresponding relationship between each spectral shape and the human motion state, and obtain motion spectrum matching data;

[0036] Among them, E z represents the energy distribution ratio of the z-th spectrum in the signal, C zl represents the complex amplitude of the z-th spectrum at the l-th frequency point, u l represents the frequency value of the l-th frequency point, and L represents the total number of frequency points.

[0037] On the other hand, the steps for obtaining the abnormal motion state index are specifically as follows:

[0038] S501: Based on the motion state classification result, calculate the change speed of the motion state in consecutive time windows, calculate the duration of the state change, compare the motion state switching frequency in different time intervals, determine the motion state fluctuation range, and generate the dynamic characteristics of the motion state;

[0039] S502: Invoke the dynamic characteristics of the motion state, analyze the fluctuation of the motion state in the short term, count the number of times of switching of the motion mode, judge the continuous stability of the motion mode, compare the signal amplitudes in the motion state change intervals, and calculate the signal energy gradient during the conversion of the motion mode to obtain the motion mode stability data;

[0040] S503: Invoke the motion mode stability data, identify the change of the signal intensity of the multi-channel echo, analyze the echo energy distribution in each time period, screen the intervals with abnormal energy fluctuations, judge the static or accelerating situation of the human body movement, and obtain the abnormal motion state index.

[0041] On the other hand, a system for recognizing the human body motion state by radar is provided. This system is applied to the method for recognizing the human body motion state by radar, and includes:

[0042] The feature extraction module uses a sensor to receive the echo signal of the millimeter-wave radar, analyzes the amplitude and phase changes of the signal, calculates the average value and standard deviation of the amplitude, compares the received signals, records the amplitude extreme points and phase mutation points, and obtains the radar echo signal features;

[0043] The motion mode analysis module, based on the radar echo signal features, segments the phase information of the signal, calculates the phase change gradient value, analyzes the gradient stable intervals, identifies the human body motion trajectory, and judges the change of the motion speed to obtain the motion mode data;

[0044] The frequency feature recognition module, based on the motion mode data, extracts the time points with abnormal signal phase and amplitude, identifies the signal frequency components, compares the frequency and energy features of the signals, judges the frequency mode of the motion state, and identifies the key frequency features to obtain the key frequency recognition results;

[0045] The spectrum matching and classification module, based on the key frequency recognition results, analyzes the spectrum shape differences, matches the known frequency modes, identifies the spectrum features during the motion, classifies the motion modes, and obtains the motion state classification results;

[0046] Based on the motion state classification result, the abnormal motion state recognition module calculates the motion change rate, analyzes the violently fluctuating motion patterns, identifies the change in the echo signal intensity, determines the abnormal human motion condition, and obtains the abnormal motion state index.

[0047] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include:

[0048] By carefully analyzing the amplitude and phase changes of the radar echo signal, the monitoring of the human motion state is optimized during the execution process. While extracting the key signal features, the amplitude average value and standard deviation of the signal are calculated for continuous time periods, further improving the sensitivity and accuracy of signal processing. In addition, by analyzing the gradient value of the phase change and comparing the changes at adjacent time points, this method effectively identifies the subtle changes in the human motion speed, greatly improving the ability to capture the dynamic motion trajectory. It not only improves the detail of signal analysis, but also enhances the adaptability and practical value of the system to complex environments through the recognition of abnormal motion states. Description of the Drawings

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0050] Figure 1 is the main step flow chart of the present invention;

[0051] Figure 2 is the system block diagram of the present invention. Detailed Embodiments

[0052] The following will describe the technical solutions in the present invention in conjunction with the drawings.

[0053] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "example" is intended to present concepts in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0054] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, their intended meanings are the same. "Of", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, their intended meanings are the same.

[0055] In the embodiments of the present invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, their intended meanings are the same.

[0056] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0057] The embodiments of the present invention provide a method for realizing human motion state recognition through radar, as Figure 1 shown, including the following steps:

[0058] S1: Use a sensor to receive the echo signal of the millimeter-wave radar, analyze the amplitude and phase changes of the echo signal, calculate the average value and standard deviation of the amplitude in a continuous time period, compare with the echo signal received in real time, screen out the significant signal change points, identify the key signal features, record the extreme points of the signal amplitude and the mutation points of the phase, and obtain the radar echo signal features;

[0059] S2: Based on the radar echo signal features, segment the phase information of the continuous echo signal, calculate the gradient value of the phase change, analyze the stable interval of the gradient change, identify the continuous human motion trajectory, compare the phase changes at adjacent time points, and judge the change of the human motion speed to obtain the motion mode data;

[0060] S3: Based on the motion mode data, extract the time points with abnormal phase and amplitude changes of the echo signal, identify the signal frequency components at the time points, compare the frequency and energy characteristics of the signals at different time points, judge the frequency mode representing different motion states, identify the key frequency features according to the energy distribution, and analyze the normal human motion state to obtain the key frequency recognition result;

[0061] S4: Based on the key frequency recognition result, analyze the spectral form differences of the signals in the motion state, call the known frequency modes, match them with each spectral form, identify the signal spectral features when the human body is moving, and classify the motion modes according to the signal features to obtain the motion state classification result;

[0062] S5: Based on the classification results of the motion states, calculate the rate of change of the motion states between adjacent windows, analyze the motion patterns with short-term severe fluctuations, identify the changes in the signal intensity of the multi-channel echoes, determine whether there are abnormal situations of human motion stillness or acceleration, and obtain the abnormal motion state indicators.

[0063] The characteristics of the radar echo signals include amplitude extreme points, phase mutation points, and signal change amplitudes. The motion pattern data includes human motion trajectories, motion speed change rates, and phase stable intervals. The key frequency identification results include dominant frequency changes, energy distribution characteristics, and frequency component differences. The classification results of the motion states include motion pattern matching degrees, spectral shape differences, and motion state categories. The abnormal motion state indicators include stillness or acceleration abnormal situations, multi-channel signal intensity changes, and motion state change rates.

[0064] The specific steps for obtaining the characteristics of the radar echo signals are as follows:

[0065] S101: Use a sensor to receive the echo signals of the millimeter-wave radar, analyze the signal amplitude and phase information, calculate the mean value and standard deviation of the signal amplitude within a continuous time period, compare the received real-time signals, identify the signal amplitude mutation intervals, calculate the signal energy change rate, and obtain the signal amplitude change characteristics;

[0066] The signal reception uses a 77GHz millimeter-wave radar module. Set the radar transmission power to 10dBm, the reception gain to 20dB, the antenna is 2 meters away from the target object. The radar module decomposes the received echo signals into in-phase components and quadrature components through an I / Q mixer, and converts the analog signals into digital signals through an analog-to-digital converter (ADC) at a sampling rate of 2.5MHz. The signal amplitude is analyzed using the amplitude calculation formula: where B(t) represents the instantaneous amplitude of the signal, X(t) represents the in-phase component of the received signal, and Y(t) represents the quadrature component of the received signal. The specific operation is as follows: During the process of receiving the echo signals, collect the signal data every 1ms, accumulate 1000 sampling points, and use the sliding window method to calculate the real-time amplitude mean value and standard deviation every 50 data points. The mean value calculation formula is: The standard deviation calculation formula is: where P is the number of sampling points, taking 1000 sampling points. The calculated mean value represents the average amplitude of the signal within 1000 sampling periods, and the standard deviation represents the degree of fluctuation of the signal amplitude. When comparing the received real-time signals, define the amplitude mutation threshold as 2 times the standard deviation of the amplitude mean value, that is, when the real-time amplitude satisfies the following conditions, it is determined as an amplitude mutation: or Among the 1000 sampling points, calculate the signal energy change rate within the amplitude mutation interval. The energy calculation uses the following formula: The energy change rate is defined as: Among them, F before represents the energy of 100 sampling points before the mutation point, and F after represents the energy of 100 sampling points after the mutation point. When the energy change rate is determined to be greater than 10%, it is a significant mutation, and the signal amplitude change characteristics are obtained.

[0067] S102: Based on the signal amplitude change characteristics, analyze the phase change of the signal, compare the phase offset between adjacent time points, analyze the continuity of the phase change, identify the cumulative offset trend of the phase change, judge the state of the echo signal, and obtain the phase offset characteristics;

[0068] The phase information of the received signal is obtained through I / Q demodulation, and the phase calculation uses the following formula: Compare the phase offset between adjacent time points, and set the phase offset as: Δθ q = θ q+1 - θ q . Within 1000 sampling periods, using the sliding window method, calculate the mean and standard deviation of the phase change every 50 data points. The mean calculation formula is: The standard deviation calculation is: Analyze the continuity of the phase change, and define the phase change mutation threshold as 2 times the standard deviation, that is: When the phase offset exceeds the mutation threshold within 5 consecutive sampling periods, it is determined as a phase mutation, identify the cumulative offset trend of the phase change, and the cumulative offset is: Define that when the cumulative offset exceeds 360 degrees, it is determined as a phase flip, determine the state of the echo signal, and obtain the phase offset characteristics.

[0069] S103: Invoke the phase offset characteristics, synchronize the signal segments with prominent signal amplitude and phase changes, calculate the change rate of the signal energy before and after the mutation point, determine the key change positions in the echo signal, analyze the extreme points of the echo signal amplitude and the phase turning points, and obtain the radar echo signal characteristics.

[0070] The mutation segments of the signal amplitude and phase are synchronized using the timestamp alignment method. Within the signal sampling period, set the window length to 100 sampling points, and define the signal energy change rate before and after the mutation point as: Among them, F before represents the energy of 100 sampling points before the mutation point, and F after represents the energy of 100 sampling points after the mutation point. Define the energy change rate threshold as 10%, that is, when the change rate satisfies the following conditions, it is determined as a mutation: K F > 0.1 or K F < -0.1, determine the key change positions in the echo signal, analyze the extreme points of the amplitude and phase, and the amplitude extreme point is defined as: Bmax = max(B(p)), B min = min(B(p)), and the phase extreme point is defined as: θ max = max(θ(q)), θ min = min(θ(q)). Determine the phase turning point, set the phase turning threshold to 30 degrees. When the phase increment of 5 consecutive data points exceeds 30 degrees, it is determined as a phase turn. Synchronously analyze the positions of the amplitude extreme point and the phase turning point to determine the characteristics of the radar echo signal.

[0071] The steps for obtaining the motion mode data are specifically as follows:

[0072] S201: Based on the characteristics of the radar echo signal, segment the phase information in the continuous echo signal, calculate the phase change amount, screen the signal intervals with abnormal phase changes, calculate the average phase increment within the interval, and compare the phase change rates of each signal segment to obtain the phase change rate index;

[0073] Calculate the average phase increment within the interval using the formula:

[0074]

[0075] And compare the phase change rates of each signal segment to obtain the phase change rate index;

[0076] Among them, represents the average phase increment within the signal interval, N represents the total number of sampling points within the selected signal interval, represents the phase increment of the i-th sampling point, represents the phase standard deviation of the i-th sampling point, represents the phase mean of the i-th sampling point;

[0077] The average phase increment within the signal interval When calculating, represents the phase increment of the i-th sampling point, and this value is obtained by discretely sampling and calculating the phase of the continuous echo signal. The calculation formula is as follows:

[0078]

[0079] Among them, and respectively represent the phase values of the i-th and the (i - 1)-th sampling points, and the phase values are obtained through the acquisition device of the radar echo signal;

[0080] The phase standard deviation The calculation formula is as follows:

[0081]

[0082] Among them, represents the phase value within a certain window interval, M represents the number of sampling points within this window, represents the average phase within this window, and the calculation formula is as follows:

[0083]

[0084] A certain radar echo signal sample collected by the data monitoring system contains N = 5 sampling points, and its phase values are respectively:

[0085] (unit: radian);

[0086] Calculate the phase increment:

[0087]

[0088] Average phase:

[0089]

[0090] Phase standard deviation:

[0091]

[0092] Calculate the denominator:

[0093]

[0094] Calculate the normalized phase increment:

[0095]

[0096] Calculate the average value of the phase increment:

[0097]

[0098] This result indicates that the average phase increment within the signal interval is 0.24875 radians, and this value is used to measure the overall phase change trend of the echo signal.

[0099] S202: Based on the phase change rate index, calculate the signal phase amplitude range within a continuous time period, compare the phase change directions of adjacent time windows, analyze the continuous change characteristics of the signal over time, identify the stable movement path of the human body, and obtain the stable movement trajectory;

[0100] Set the signal phase change to θ(t), where t represents time, collect signal data at a sampling rate of 1000 Hz, set the length of a single time window to 100 ms, each time window contains 100 sampling points, calculate the phase amplitude range through the sliding window method, and the phase amplitude range is defined as the difference between the maximum phase and the minimum phase within the current window. The calculation formula is: Δθamp = θ max (t) - θ min (t), where θ max (t) and θ min (t) respectively represent the maximum phase and the minimum phase within the current window. Assuming the phase data within the current window is: θ = [10°, 12°, 15°, 9°, 8°, 14°, 11°, 13°, 16°, 17°], then the maximum phase is θ max (t) = 17°, and the minimum phase is θ min (t) = 8°. The phase amplitude range is: Δθ amp = 17° - 8° = 9°. Comparing the phase change directions of adjacent time windows, the phase change direction is defined as: D θ (t) = sign(θ t+1 - θ t ), where the sign function is defined as: Assuming the current adjacent phase data sequence is [10°, 12°, 15°, 9°, 8°], then the phase change direction is: D θ (t) = [1, 1, -1, -1]. Analyzing the continuous change characteristics of the signal over time, if the phase change direction remains the same within 5 consecutive time windows, it is defined as a stable phase change state. If the phase change direction alternates within 3 consecutive time windows, it is defined as an unstable phase change state. Judging the trajectory change trend in the stable state, with a 5 - time - window period, accumulating the phase change trend: If Θ trend > 3 or Θ trend < - 3, it is defined as a significant trend change, identifying the stable movement path of the human body to obtain a stable movement trajectory.

[0101] S203: Invoke the stable movement trajectory, analyze the direction of human movement, screen the time nodes of the change in the movement trajectory direction, judge the amplitude and duration of the trajectory direction adjustment, compare the movement speeds of adjacent trajectory segments, identify the change in the movement rhythm of the human body, record the change intervals of the movement direction and speed, and obtain the movement pattern data.

[0102] Define the phase change direction within the current time window as: D θ (t) = sign(θ t+1 - θ t ). Define the amplitude of the trajectory direction adjustment as the phase difference between adjacent time windows. The calculation formula is: Δθ adjust = θ next - θ current, assume that the data sequence of the current trajectory segment is: θ = [10°, 12°, 15°, 9°, 8°, 14°, 11°, 13°, 16°, 17°]. Between the 3rd and 4th time windows, the phase adjustment amplitude is: Δθ adjust = 9° - 15° = -6°, and the duration of the direction adjustment is the number of adjacent window intervals, defined as: T θ = n × Δt, where n is the number of windows and Δt is the length of each window (100 ms). If the duration of the direction adjustment is 3 windows, then: T θ = 3 × 100 ms = 300 ms. Comparing the motion speeds of adjacent trajectory segments, the motion speed is defined as the phase change amount per unit time. The speed calculation formula is: Substitute the data: If the speed change of adjacent trajectory segments exceeds 20%, it is defined as a sudden change in motion speed. The rate of speed change is defined as: Assume that the speeds of adjacent trajectory segments are [-20° / s, -24° / s], then the rate of speed change is: If the rate of change is greater than 20%, it is determined as a significant speed change, and the change intervals of the motion direction and speed are recorded to obtain the motion mode data.

[0103] The specific steps for obtaining the key frequency identification results are as follows:

[0104] S301: Based on the motion mode data, detect the phase and amplitude fluctuations of the echo signal, screen out the time points with prominent change rates, calculate the amplitude difference of the continuous signals before and after the time points, analyze the change law of the phase mutation, measure the duration of the signal fluctuation, screen out the signal feature points of the amplitude and phase mutations, and obtain the signal fluctuation feature set;

[0105] The received echo signal is resolved into the in-phase component U(τ) and the quadrature component V(τ) through an I / Q demodulator. The phase calculation formula is: The amplitude calculation formula is: Screen out the time points with prominent change rates, and set the phase fluctuation rate as: The amplitude fluctuation rate is: Among them, Γτ is the time interval between adjacent sampling points, with a value of 1 ms. Set the mutation threshold as the mean of the first 100 data points plus 2 times the standard deviation. The mutation threshold formula is defined as: Among them, and are the means of the first 100 data points, and δ ψ and δ W are the standard deviations. When the following conditions are met, it is determined as a mutation point: or Calculate the amplitude difference of the continuous signal before and after the time point. The amplitude difference is calculated as: ΓW = W after - W before , analyze the variation law of the phase mutation, and define the phase mutation rate as: Measure the duration of the signal fluctuation. The fluctuation duration is defined as the time interval from the mutation point to the restored stable state. The stable state is defined as the time point when the signal returns to within the range of the mean ± 1 times the standard deviation. Screen the signal feature points of amplitude and phase mutation to obtain the signal fluctuation feature set.

[0106] S302: Call the signal fluctuation feature set, extract the echo signal at the corresponding time point, analyze the amplitude change of the signal in the different frequency bands, identify the frequency components in the signal, calculate the rate of change of frequency over time, judge the frequency offset trend between time points, and analyze the frequency distribution in adjacent time windows to obtain the frequency offset spectrum;

[0107] Use the short-time Fourier transform (STFT) to analyze the amplitude change of the signal in the different frequency bands. The STFT transform formula is: where G(g,τ) represents the complex form of the time-frequency signal, w(γ - τ) is the window function. The Hanning window (HanningWindow) is selected, the window length is set to 100 ms, and the overlap rate is set to 50%. Analyze the amplitude change of the signal in the different frequency bands, extract the amplitude of the frequency components at an interval of 1 Hz frequency increment, identify the frequency components in the signal, and define the frequency change rate as: Judge the frequency offset trend between time points. The determination condition of the frequency offset trend is that the frequency change directions are the same in 5 consecutive windows. If the following conditions are met, it is defined as a frequency offset trend mutation: Analyze the frequency distribution in adjacent time windows to obtain the frequency offset spectrum.

[0108] S303: Based on the frequency offset spectrum, analyze the energy density in each frequency interval, compare the energy proportion of each frequency band, screen the stable frequency region, identify the frequency characteristics in the normal motion state, judge the matching relationship between the motion states of the key frequency modes, and obtain the key frequency identification result.

[0109] Set the energy density calculation formula as: Compare the energy proportion of each frequency band, and define the energy proportion of the frequency band as: Screen the stable frequency region. The stable frequency interval is defined as the frequency interval where the frequency change amplitude is lower than 2 Hz and the energy proportion exceeds 20% within 5 consecutive time windows. Identify the frequency characteristics in the normal motion state, judge the matching relationship between the motion states of the key frequency patterns, set the key frequency threshold range as the frequency components between 40 Hz and 70 Hz, and determine that when the energy proportion within this frequency interval exceeds 30%, it is the key frequency pattern, and obtain the key frequency recognition result.

[0110] The steps to obtain the motion state classification result are specifically as follows:

[0111] S401: Based on the key frequency recognition data, analyze the spectral morphology of the signal during human motion, extract the signal characteristics corresponding to each frequency pattern, compare the morphological differences of the signal waveforms, identify the spectral changes of each motion state, and obtain the spectral difference recognition result;

[0112] The received signal extracts frequency characteristics through the short-time Fourier transform (STFT). Define the spectral morphology of the signal in the time ξ and frequency f dimensions as: P(f, ξ) = |F(f, ξ)| 2 , where P(f, ξ) represents the power spectral density at time ξ and frequency f, and F(f, ξ) represents the complex spectrum of the signal. Extract the signal characteristics corresponding to each frequency pattern, and define the feature extraction formula as: where, E k represents the eigenvalue of the k-th frequency pattern, and W k (f) represents the weight function of the k-th frequency pattern, and its value is the normalized Gaussian distribution function, defined as: where, μ k is the central frequency of the k-th frequency pattern, and σ k is the standard deviation of the k-th frequency pattern. Set the central frequency value range as 40 Hz to 70 Hz, and the standard deviation as 5 Hz. The extracted frequency pattern eigenvalues are processed by the principal component analysis (PCA) method for dimensionality reduction, retaining more than 95% of the feature information. Compare the waveform morphological differences under different frequency patterns, and define the morphological difference degree as: where, P k (ξ) represents the power spectral density of the k-th frequency pattern at time ξ, represents the average power spectral density of the k-th frequency pattern. Compare the morphological difference degrees of different patterns, screen out the patterns with a morphological difference degree exceeding 1.5 times the standard deviation, and determine them as significant spectral differences to obtain the spectral difference recognition result.

[0113] S402: Invoke the spectrum difference recognition result, compare with the known frequency patterns, filter out the spectrum forms that conform to the pattern characteristics, calculate the energy distribution ratio of each spectrum in the signal, analyze the matching degree between each pattern, determine the corresponding relationship between each spectrum form and the human body movement state, and obtain the motion spectrum matching data;

[0114] Calculate the energy distribution ratio of each spectrum in the signal, using the formula:

[0115]

[0116] Analyze the matching degree between each pattern, determine the corresponding relationship between each spectrum form and the human body movement state, and obtain the motion spectrum matching data;

[0117] where, E z represents the energy distribution ratio of the z-th spectrum in the signal, C zl represents the complex amplitude of the z-th spectrum at the l-th frequency point, u l represents the frequency value of the l-th frequency point, and L represents the total number of frequency points;

[0118] C zl represents the complex amplitude of the z-th spectrum at the l-th frequency point, which can be obtained through the fast Fourier transform (FFT) of the signal. The specific calculation process is as follows:

[0119] Perform Fourier transform on the time-domain signal x(t) to obtain the complex spectrum expressed as:

[0120]

[0121] The complex amplitude is calculated as:

[0122]

[0123] where:

[0124] x(t) is the sampled time-domain signal, and the sampling frequency is set to 1000 Hz, which is directly obtained through a high-precision sensor;

[0125] f l is the l-th frequency point, with a value range of 0 Hz to 500 Hz, a frequency interval of 1 Hz, and 500 frequency points are taken;

[0126] Re(X(f l )) is the real part of the spectrum at the l-th frequency point;

[0127] Im(X(f l )) is the imaginary part of the spectrum at the l-th frequency point;

[0128] u lThe frequency value representing the $l$-th frequency point is directly output by the spectrum analysis system. The frequency resolution is 1 Hz, the frequency range is from 0 Hz to 500 Hz, and the number of frequencies $L = 500$.

[0129] Based on the above content, the parameters are set as follows:

[0130] The frequency range corresponding to the $z$-th spectrum is from 20 Hz to 100 Hz, and the complex amplitude is obtained in the following way:

[0131] The complex amplitude at the 20 Hz frequency point

[0132] The complex amplitude at the 50 Hz frequency point

[0133] The complex amplitude at the 100 Hz frequency point

[0134] The frequency values are set as follows:

[0135] The frequency value $u$ at 20 Hz 20 $= 20$, the frequency value $u$ at 50 Hz 50 $= 50$, the frequency value $u$ at 100 Hz 100 $= 100$;

[0136] Substitute the above parameters into the formula:

[0137]

[0138] Perform specific calculations:

[0139]

[0140] Expand the denominator term for calculation:

[0141] 1.0 2 + 1.28 2 + 0.78 2 $= 1.0 + 1.6384 + 0.6084 = 3.2468$;

[0142] 20 2 + 50 2 + 100 2 $= 400 + 2500 + 10000 = 12900$;

[0143] Fully expand the denominator:

[0144]

[0145] Expand the numerator term:

[0146] $20 + 64 + 78 = 162$;

[0147] The calculation result is:

[0148]

[0149] This result indicates that the energy distribution ratio of the z-th spectrum in the signal is 0.791, reflecting the proportion of the z-th spectrum form in the overall signal.

[0150] S403: Call the motion spectrum matching data, analyze the proportion of the motion pattern signal, screen out the significantly proportioned motion patterns, analyze the time persistence of the corresponding patterns, compare the signal frequency changes between different motion states, determine the motion category to which the echo signal belongs, determine the motion pattern classification interval, and obtain the motion state classification result.

[0151] Call the motion spectrum matching data, analyze the proportion of the motion pattern signal, and define the proportion of the motion pattern signal as: where J k represents the signal energy proportion of the k-th frequency pattern, and P k (f, ξ) represents the power spectral density of the k-th frequency pattern at time ξ and frequency f. Screen out the significantly proportioned motion patterns, define the significant threshold as 0.2 (i.e., the proportion exceeds 20%), and if the following conditions are met, it is determined as a significant motion pattern: J k > 0.2. Analyze the time persistence of the corresponding pattern, define the time persistence as continuously exceeding 5 time windows, and each window length is 100 ms. The time persistence is calculated as: T k = n × Δξ, where n is the number of time windows and Δξ is the window length (100 ms). If the time persistence exceeds 500 ms, it is determined as a stable motion pattern. Compare the signal frequency changes between different motion states, and define the frequency change between adjacent patterns as: Δf k = f next - f current , and define the frequency change rate between adjacent patterns as: If the frequency change rate exceeds 10 Hz / s, it is determined as a significant frequency change. Determine the motion category to which the echo signal belongs, define the motion pattern classification interval as: 40 Hz to 50 Hz → slow walking, 50 Hz to 60 Hz → normal walking, 60 Hz to 70 Hz → fast walking. Match the current frequency change pattern with the above classification interval. If the corresponding frequency interval and frequency change rate conditions are met, determine the motion category to which the current echo signal belongs, determine the motion pattern classification interval, and obtain the motion state classification result.

[0152] The steps for obtaining the abnormal motion state index are specifically as follows:

[0153] S501: Based on the classification result of the motion state, calculate the change speed of the motion state in consecutive time windows, calculate the duration of the state change, compare the motion state switching frequency in the differential time, judge the motion state fluctuation range, and generate the dynamic characteristics of the motion state;

[0154] Define the motion state change as the conversion of the motion mode within the time window, set the time window length to 100 ms, with 10 time windows per second, and define the state change speed as: where V k represents the motion state change speed, N c represents the number of motion mode changes that occur within the window length, Q w represents the window length (100 ms). Assuming that 5 mode switches occur within a 1-second window, then: times per second. Calculate the duration of the state change. Define the duration of the state change as the time difference between two adjacent motion modes, and the calculation formula is: Ω k = t2 - t1, where t2 and t1 respectively represent the time points of two adjacent motion mode switches. Assuming that two adjacent modes occur at 500 ms and 1200 ms, then: Ω k = 1200 ms - 500 ms = 700 ms. Compare the motion state switching frequency in the differential time. Define the switching frequency as the number of switches that occur within the unit time (second), and the formula is: where Q is the sampling interval length. Assuming that 10 mode switches occur within 2 seconds, then: times per second. Judge the motion state fluctuation range. Define the fluctuation range as the time length between two adjacent state changes. When the fluctuation range is less than 500 ms and there are 3 or more consecutive state switches, it is defined as a high-frequency motion fluctuation range, and generate the dynamic characteristics of the motion state.

[0155] S502: Call the dynamic characteristics of the motion state, analyze the fluctuation of the motion state in the short term, count the number of switches of the motion mode, judge the continuous stability of the motion mode, compare the signal amplitude in the motion state change interval, calculate the signal energy gradient when the motion mode is converted, and obtain the motion mode stability data;

[0156] Define the fluctuation degree as the number of state changes that occur within a short time window (500 ms). Assuming that 3 state switches occur within a 500-ms window, define the short-term fluctuation degree as: times per second. Count the number of switches of the motion mode. Define the number of switches as the total number of state switches that occur within the specified time window. Assuming that 25 mode switches occur within 10 seconds, then: N c= 25, determine the continuous stability of the motion mode. Define stability as the proportion of a certain state in the total time within a period of time, and define it as: where T stable represents the duration in this mode, and T total represents the total time. Assuming that the duration of this mode is 7 seconds within 10 seconds, then: Compare the signal amplitudes in the motion state change interval, and set the signal amplitude difference as: Ω N = N next - N current where N next and N current are the signal amplitudes in adjacent windows respectively. Assuming that the signal amplitude in the current window is 0.8 and the next window is 1.2, then: Ω N = 1.2 - 0.8 = 0.4. Calculate the signal energy gradient during the motion mode conversion. Define the energy gradient as the rate of energy change in adjacent windows, and the calculation formula is: where L next and L current are the energies of adjacent windows respectively. Assuming that the energy in the current window is 2.5 and the next window is 3.0, and the window length is 100 ms, then: (unit / s), and obtain the motion mode stability data.

[0157] S503: Call the motion mode stability data, identify the signal intensity change of the multi-channel echo, analyze the echo energy distribution in each time period, screen the intervals with abnormal energy fluctuations, judge the static or accelerating situation of human motion, and obtain the abnormal motion state index.

[0158] Call the motion mode stability data, identify the signal intensity change of the multi-channel echo, and set the signal intensity of the multi-channel received echo as: where I m (ζ) represents the signal intensity of the m-th receiving channel at time ζ, and R m (ζ) and S m (ζ) are the in-phase component and the quadrature component respectively. Analyze the echo energy distribution in each time period, and define the energy as: Screen the intervals with abnormal energy fluctuations, and define the abnormal fluctuation threshold as the mean value of the energy plus 2 times the standard deviation: where is the mean value of the energy, and δ L is the standard deviation of the energy. If the energy of a certain channel exceeds the threshold, it is determined that the energy fluctuation is abnormal, judge the static or accelerating situation of human motion, and obtain the abnormal motion state index.

[0159] As Figure 2 shown, the system for realizing human motion state recognition through radar includes:

[0160] The feature extraction module uses a sensor to receive the echo signal of the millimeter-wave radar, analyzes the amplitude and phase changes of the signal, calculates the average value and standard deviation of the amplitude, compares the received signal, records the amplitude extreme points and phase mutation points, and obtains the radar echo signal features;

[0161] The motion mode analysis module, based on the radar echo signal features, segments the phase information of the signal, calculates the phase change gradient value, analyzes the gradient stable interval, identifies the human body motion trajectory, judges the change of the motion speed, and obtains the motion mode data;

[0162] The frequency feature recognition module, based on the motion mode data, extracts the time points with abnormal signal phase and amplitude, identifies the signal frequency components, compares the frequency and energy features of the signal, judges the frequency mode of the motion state, identifies the key frequency features, and obtains the key frequency recognition results;

[0163] The spectrum matching and classification module, based on the key frequency recognition results, analyzes the spectrum shape differences, matches the known frequency modes, identifies the spectrum features during motion, classifies the motion modes, and obtains the motion state classification results;

[0164] The abnormal motion state recognition module, based on the motion state classification results, calculates the motion change rate, analyzes the violently fluctuating motion modes, identifies the change of the echo signal intensity, judges the abnormal human body motion conditions, and obtains the abnormal motion state indicators.

[0165] It should be understood that the term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Among them, A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context.

[0166] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one (item)" or its similar expression means any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0167] It should be understood that in various embodiments of the present invention, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0168] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0169] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described devices, apparatuses, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0170] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0171] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0172] In addition, in each embodiment of the present invention, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0173] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0174] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for recognizing human motion states through radar, characterized in that The method includes: S1: Use a sensor to receive the echo signal of the millimeter-wave radar, analyze the amplitude and phase changes of the signal, calculate the average value and standard deviation of the amplitude, compare the received signal, record the amplitude extreme points and phase mutation points, and obtain the radar echo signal characteristics; S2: Based on the radar echo signal characteristics, segment the phase information of the signal, calculate the phase change gradient value, analyze the gradient stable interval, identify the human body movement trajectory, judge the change of the movement speed, and obtain the movement mode data; S3: Based on the movement mode data, extract the time points with abnormal signal phase and amplitude, identify the signal frequency components, compare the frequency and energy characteristics of the signal, judge the frequency mode of the movement state, identify the key frequency characteristics, and obtain the key frequency identification result; S4: Based on the key frequency identification result, analyze the spectral morphology difference, match the known frequency mode, identify the spectral characteristics during movement, classify the movement mode, and obtain the movement state classification result; S5: Based on the movement state classification result, calculate the movement change rate, analyze the movement mode with severe fluctuations, identify the change of the echo signal intensity, judge the abnormal situation of the human body movement, and obtain the abnormal movement state index.

2. The method for recognizing human motion state through radar according to claim 1, wherein, The radar echo signal characteristics include amplitude extreme points, phase mutation points and signal change amplitude. The movement mode data includes the human body movement trajectory, movement speed change rate and phase stable interval. The key frequency identification result includes dominant frequency change, energy distribution characteristics and frequency component difference. The movement state classification result includes movement mode matching degree, spectral morphology difference and movement state category. The abnormal movement state index includes static or acceleration abnormal situation, multi-channel signal intensity change, movement state change rate.

3. The method for identifying the human body motion state through radar according to claim 1, characterized in that, The specific steps for obtaining the radar echo signal characteristics are as follows: S101: Use a sensor to receive the echo signal of the millimeter-wave radar, analyze the signal amplitude and phase information, calculate the average value and standard deviation of the signal amplitude in a continuous time period, compare the received real-time signal, identify the signal amplitude mutation interval, calculate the signal energy change rate, and obtain the signal amplitude change characteristics; S102: Based on the signal amplitude change characteristics, analyze the phase change of the signal, compare the phase offset amount of adjacent time points, analyze the continuity of the phase change, identify the cumulative offset trend of the phase change, judge the echo signal state, and obtain the phase offset characteristics; S103: Call the phase offset characteristics, synchronize the signal segments with prominent signal amplitude and phase changes, calculate the change rate of the signal energy before and after the mutation point, determine the key change positions in the echo signal, analyze the amplitude extreme points and phase turning points of the echo signal, and obtain the radar echo signal characteristics.

4. The method for identifying the human body motion state through radar according to claim 1, characterized in that The specific steps for obtaining the movement mode data are as follows: S201: Based on the radar echo signal characteristics, segment the phase information in the continuous echo signal, calculate the phase change amount, screen the signal intervals with abnormal phase changes, calculate the average value of the phase increment in the interval, and compare the phase change rate of each signal segment to obtain the phase change rate index; S202: Based on the phase change rate index, calculate the signal phase amplitude range within a continuous time period, compare the phase change directions of adjacent time windows, analyze the continuous change characteristics of the signal over time, identify the stable movement path of the human body, and obtain the stable movement trajectory; S203: Invoke the stable movement trajectory, analyze the direction of human movement, screen the time nodes where the movement trajectory direction changes, judge the amplitude and duration of the trajectory direction adjustment, compare the movement speeds of adjacent trajectory segments, identify the change in the movement rhythm of the human body, record the change intervals of the movement direction and speed, and obtain the movement mode data.

5. The method for identifying the human body movement state through radar according to claim 4, wherein The average phase increment within the calculation interval adopts the formula: And compare the phase change rates of each signal segment to obtain the phase change rate index; Among them, represents the average phase increment within the signal interval, N represents the total number of sampling points within the selected signal interval, represents the phase increment of the i-th sampling point, represents the phase standard deviation of the i-th sampling point, represents the phase mean of the i-th sampling point.

6. The method for identifying the human body movement state through radar according to claim 1, characterized in that, The steps for obtaining the key frequency recognition result are specifically as follows: S301: Based on the movement mode data, detect the phase and amplitude fluctuations of the echo signal, screen the time points with prominent change rates, calculate the amplitude difference of the continuous signals before and after the time points, analyze the change law of the phase mutation, measure the duration of the signal fluctuation, screen the signal feature points of the amplitude and phase mutation, and obtain the signal fluctuation feature set; S302: Invoke the signal fluctuation feature set, extract the echo signal at the corresponding time points, analyze the amplitude change of the signal in different frequency bands, identify the frequency components in the signal, calculate the change rate of the frequency over time, judge the frequency offset trend between the time points, and analyze the frequency distribution in adjacent time windows to obtain the frequency offset map; S303: Based on the frequency offset map, analyze the energy density of each frequency interval, compare the energy ratios of each frequency band, screen the stable frequency regions, identify the frequency characteristics in the normal movement state, judge the matching relationship between the key frequency modes and the movement states, and obtain the key frequency recognition result.

7. The method for identifying human motion state through radar according to claim 1, characterized in that The steps for obtaining the movement state classification result are specifically as follows: S401: Based on the key frequency recognition data, analyze the spectral shape of the signal during human movement, extract the signal characteristics corresponding to each frequency mode, compare the morphological differences of the signal waveforms, identify the spectral changes of each movement state, and obtain the spectral difference recognition result; S402: Invoke the spectral difference recognition result, compare the known frequency modes, screen the spectral shapes that conform to the mode characteristics, calculate the energy distribution ratio of each spectrum in the signal, analyze the matching degree between each mode, determine the corresponding relationship between each spectral shape and the human movement state, and obtain the movement spectrum matching data; S403: Invoke the movement spectrum matching data, analyze the proportion of the movement mode signals, screen the movement modes with significant proportions, analyze the time persistence of the corresponding modes, compare the signal frequency changes between different movement states, judge the movement category to which the echo signal belongs, determine the movement mode classification interval, and obtain the movement state classification result.

8. The method for recognizing human motion state by radar according to claim 1, wherein The formula for calculating the energy distribution ratio of each spectrum in the signal is: Analyze the matching degree between each mode, determine the corresponding relationship between each spectral shape and the human movement state, and obtain the movement spectrum matching data; Among them, E z represents the energy distribution ratio of the z-th spectrum in the signal, C zl represents the complex amplitude of the z-th spectrum at the l-th frequency point, u l represents the frequency value of the l-th frequency point, and L represents the total number of frequency points.

9. The method for identifying human motion state through radar according to claim 1, characterized in that The steps for obtaining the abnormal movement state index are specifically as follows: S501: Based on the classified results of the motion states, calculate the change speed of the motion states in a continuous time window, calculate the duration of the state change, compare the motion state switching frequencies at different times, determine the motion state fluctuation range, and generate dynamic features of the motion states. S502: Invoke the dynamic features of the motion states, analyze the fluctuation of the motion states in the short term, count the number of switches of the motion patterns, judge the continuous stability of the motion patterns, compare the signal amplitudes in the motion state change range, and calculate the signal energy gradient during the conversion of the motion patterns to obtain the motion pattern stability data. S503: Invoke the motion pattern stability data, identify the change of the signal intensity of the multi-channel echoes, analyze the echo energy distribution in each time period, filter out the intervals with abnormal energy fluctuations, judge the static or accelerating situation of the human body motion, and obtain the abnormal motion state index.

10. A system for recognizing human motion states through radar, the system for recognizing human motion states through radar being used to implement the method for recognizing human motion states through radar according to any one of claims 1-9, characterized in that, The system includes: The feature extraction module uses a sensor to receive the echo signal of the millimeter-wave radar, analyzes the amplitude and phase changes of the signal, calculates the average value and standard deviation of the amplitude, compares the received signals, and records the amplitude extreme points and phase mutation points to obtain the radar echo signal features. The motion pattern analysis module, based on the radar echo signal features, segments the phase information of the signal, calculates the phase change gradient value, analyzes the gradient stable interval, identifies the human body motion trajectory, and judges the change of the motion speed to obtain the motion pattern data. The frequency feature recognition module, based on the motion pattern data, extracts the time points with abnormal signal phase and amplitude, identifies the signal frequency components, compares the frequency and energy features of the signals, judges the frequency pattern of the motion states, and identifies the key frequency features to obtain the key frequency recognition results. The spectrum matching and classification module, based on the key frequency recognition results, analyzes the spectrum shape differences, matches the known frequency patterns, identifies the spectrum features during the motion, classifies the motion patterns, and obtains the classified results of the motion states. The abnormal motion state recognition module, based on the classified results of the motion states, calculates the motion change rate, analyzes the violently fluctuating motion patterns, identifies the change of the echo signal intensity, judges the abnormal situation of the human body motion, and obtains the abnormal motion state index.

Images